Numerous drugs are described in scientific literature as causative of TMA, mostly as little case series, and the number is rapidly increasing (see Table 1). The drugs reported as causative of TMA are heterogeneous in terms of mechanism of action. Some drugs are more frequently used in clinical activity, as antibiotics or antiplatelet therapy, some others in more restricted contexts, such as IO-anti-VEGF. The heterogeneity of these causative drugs should require that physicians from several different areas are familiar with DITMA.

The most frequent drugs described as TMA causative agents are: calcineurin inhibitors (tacrolimus, cyclosporine, etc), thienopyridine (clopidogrel), anti-neoplastic drugs (gemcitabine, mitomycin, etc), VEGF inhibitors (bevacizumab) and VEGF receptor blockers (tyrosine kinase inhibitors, such as pazopanib and sunitinib), interferon (alpha, beta, gamma) and quinine. These groups of drugs account for more than 75% of all DITMA cases (Reese et al., 2015; Bayer et al., 2019).

Considering that DITMA can be caused by different and heterogeneous mechanisms (Figure 1), the most frequent etiopathogenetic classification splits them in two subgroups: immune-mediated (idiosyncratic) and dose-related/toxic.

1) The idiosyncratic immune-mediated DITMA is related to abnormal susceptibility/hypersensitivity to a given drug. It is characterized by the production of reactive antibodies after the exposure to the putative drug. The antibodies interact with endothelium, platelets and/or circulating factors, favoring platelet-thrombi formation and endothelial damage, thus predisposing to the TMA development. The reported causative antibodies are mainly directed against platelets. Drug-dependent antiplatelet antibodies have been demonstrated for quinine (Al-Nouri et al., 2015; George et al., 2017), vancomycin and oxaliplatin (Reese et al., 2015). Moreover, other autoantibodies are reported but not clearly associated with drug exposure. For example, the presence of drug-dependent antibodies against ADAMTS-13 together with decreased anti-ADAMTS-13 activity are reported for interferon (Orvain et al., 2014; Gangaraju et al., 2017; Kundra and Wang, 2017) and for ticlopidine (Tsai et al., 2000; Zakarija et al., 2009; Bennett et al., 2013). Evaluating those case reports, the association between drug exposure, drug-dependent anti-ADAMTS-13 antibody and TMA did not fit with the temporal criteria adopted to define an immune-mediated DITMA. In conclusion, based on the literature data, it is possible to hypothesize the presence of drug-dependent anti-ADAMTS-13 antibodies as a causative mechanism underlying a few cases of immune-mediated DITMA (Figure 1), but further supporting evidence are needed to prove their causative role.

This second group has not a clear time onset after drug exposure (classically the TMA occurs after months) and no adding criteria for DITMA diagnosis is needed.

This classification has some limitations. First of all, some drugs can cause TMA both through immune effects and directly. This is described for example for interferon (Kavanagh et al., 2016), oxaliplatin and gemcitabine (Al-Nouri et al., 2015; Saleem et al., 2018). Second, the toxic group comprises different drugs, having heterogeneous biological targets, such as signal transducer, transcription factor or receptors. Moreover, these targets have different distributions above tissues and organs. This target heterogeneity is reflected in clinical DITMA manifestation. In fact, as described above, anti-VEGF and TK inhibitors have tropism for endothelial-podocyte complexes. The deregulation of the podocyte-endothelium complex results in a particular clinical entity described as renal-limited TMA (Izzedine and Perazella, 2015). In contrast to systemic TMA, renal-limited TMA is defined as biopsy-proven renal TMA in the absence of microangiopathic hemolytic anemia and thrombocytopenia. Renal-limited DITMA might be underdiagnosed due to clinical findings limited to the kidney and physicians’ frequent reluctance to perform renal biopsy. This issue becomes even more important in DITMA, considering that TMA renal-limited forms are mainly caused by drugs (28.5%) (Saba et al., 2018). Moreover, if recognized, these forms showed to benefit more from the withdrawal of the causative drug, with a lower relapse risk and better survival rates than the systemic DITMA (Izzedine and Perazella, 2015). These considerations emphasize the importance of considering separately kidney-limited and systemic DITMA forms.

Moreover, some drugs are not able to cause TMA when administered alone but were reported to induce TMA when administered together with other drugs. This includes cytotoxic agents, such as bleomycin and cisplatin (Salhi et al., 2021), and immune checkpoint inhibitors, such as ipilimumab and nivolumab (Lafranchi et al., 2020). In consideration of the criteria listed above for the definition of causative association, we cannot report these drugs as “causative” for TMA. Nevertheless, a synergic effect between endothelial damage, ADAMTS13 deficiency and complement hyperactivation in the development of microangiopathy has already been hypothesized for TMA and demonstrated in vivo models (Zheng et al., 2019). There is the possibility that this mechanism is associated with DITMA, but further studies are needed to investigate this hypothesis.

The initial trigger for endothelial cells damage could involve i) platelets, ii) ADAMTS13-activity, iii) endothelial cells, iv) endothelial-podocyte complex. One of the final effectors of endothelial damage is the complement system (Mathern and Heeger, 2015). Mutations of complement proteins such as CFH, C3, CFI, CFB or MCP are clearly associated with development of primary TMA with deregulated activity of the complement system (Noris et al., 2010; Nester et al., 2015). Among secondary TMA forms, the role of the complement is more controversial and difficult to clarify (Bresin et al., 2013; Fremeaux-Bacchi et al., 2013) considering, for example the pan-reactivity of different potential TMA-triggers such as infections (Kobbe et al., 2017), organ and marrow transplantation (Alasfar and Alachkar, 2014), pregnancy (Fakhouri et al., 2010) and, more rarely, drugs (Noris and Remuzzi, 2015). A clear association between complement activation and endothelial damage has not been demonstrated for DITMA. Complement involvement is partially documented in secondary TMAs by high levels of activation products (C3a and sC5b-9), low CFH levels (Farkas et al., 2017), low complement C3 levels (Gilbert et al., 2013), elevated BB fragment (Bhutani et al., 2020; Taghavi et al., 2022), and EC’s deposition of C5b-C9 complex in EC-culture and nephron tissue (Palomo et al., 2019; Blasco et al., 2021). These laboratory/histological data indicate an upregulation of the complement system in TMA.

We performed a literature review in order to understand the extent of the complement activation searching both for decreased serum value of complement proteins and complement deposition in kidney biopsy. More than half of the patients with DITMA who underwent a renal biopsy showed the presence of complement factors at IF staining (38/66; 59%) (Table 2), mainly identified in the capillary wall (24/27; 88%). Among these 38 patients, 25 (66%) have a concomitant evaluation of serum complement factors available which showed normal serum complement levels in 16/25 (64%) patients and decreased complement levels in 9/25 (36%) patients (mainly characterized by with isolated C3 reduction).

Focusing on those patients with decreased serum complement factors, they all had positive IF staining. This concordance may help to identify, based on simple, inexpensive and widely available blood tests, a cluster of patients with an underlying complement hyperactivation. It seems reasonable to treat with anti-complementary therapy, particularly those patients with signs of complement activation.

Different drugs seem to mediate different complement involvement: DITMA cases related to anti-TK (6/7, 95%) and intraocular anti-VEGF (IO anti-VEGF) (5/5, 100%) drugs have a negative complement IF in most of the cases, while DITMA cases secondary to gemcitabine (14/15, 97%) and ramucirumab (11/11, 100%) present complement activation in almost all cases.

These data suggest a variable complement involvement, a likely different susceptibility to anti-complementary drugs and an heterogeneity of the DITMA mechanisms. This will require deeper evaluation of the DITMA pathophysiological mechanism for future reclassifications and drug/disease-based treatment. Nevertheless, several problems still need to be solved to clarify the role of complement in DITMA (as in other secondary TMA forms).

The detection of mutations in complement genes of patients with secondary TMA (Bruel et al., 2017; Gaggl et al., 2018; Palma et al., 2021) extended the concept of “deregulated” complement system at the molecular level not only in primary forms of TMA. Some observations highlighted the pivotal role of genetic susceptibility in endothelial damage development; for example, TMA haplotype risks (Esparza-Gordillo et al., 2005; Ding et al., 2017) are more frequent in secondary TMAs (Le Clech et al., 2019) than in healthy samples. Nevertheless, there are suggestions of no genetic involvement (Le Clech et al., 2019; Fakhouri and Frémeaux-Bacchi, 2021) in the genesis of endothelial damage, in favor of single (40% of cases) or multiple (60% of cases) environmental triggers (Bayer et al., 2019).

Genetic variants in complement genes are found in less than 10% of DITMA cases (Cavero et al., 2017; Palma et al., 2021), and only a small number of case reports is reported on the detection of potential causative complement variants in DITMA (Gilbert et al., 2013; Atallah-Yunes and Soe, 2018).

To the best of our knowledge, in literature, the total number of DITMA patients with available genetic panels for the screening of complement genes is 122 (Table 3). The whole exome sequencing was not carried out in any case. Pathogenic variants (4–5 class in ACMG classification) were reported in only two patients (2/122, 1.7%) and variants of unknown significance (VOUS) or class 3 in ACMG classification were found in 12/98 (12.2%) also considering CFHR1-3 deletions. However, as suggested by recent data (Le Clech et al., 2019; Fakhouri and Frémeaux-Bacchi, 2021), those variants are not clearly associated with TMA susceptibility and should not be considered clear genetic risk factors for TMA. This lowers the total number of patients with VOUS to 9/98 (9.1%). In the general population, VOUS in complement genes (mainly CFH and CFHR) are reported in 5%–10% of healthy controls (Le Clech et al., 2019; Fakhouri and Frémeaux-Bacchi, 2021). The percentage of VOUS in DITMA cases is not significantly different from that of the general population. Taking these data into account, the role of pathogenic variants in complement genes in DITMA is probably limited to a few exceptional cases (2/122, 1.7% of patients in literature with likely pathogenic or pathogenic variant), differently from what happened in aHUS, malignant hypertension-associated TMA (Zhang et al., 2020) and pregnancy-related TMA (Frémeaux-Bacchi et al., 2019). In support of this, the DITMAs showed a very low relapse rate (Le Clech et al., 2019).

The absence of genetic abnormalities in most patients does not exclude participation of complement in the pathogenesis of DITMA and other secondary TMAs. Complement dysregulation is not equivalent to complement activation. In fact, the activated procoagulant and proinflammatory endothelial cell phenotype characteristic of DITMA can induce a non-specific activation of complement, which in turn can promote and aggravate TMA (Duineveld and Wetzels, 2019; Frémeaux-Bacchi et al., 2019).

Therapeutic plasma exchange (TPE) (Zbaras et al., 2013; Cepeda et al., 2018) and rituximab (RTX) (Gourley et al., 2010; Murugapandian et al., 2015) have been utilized among patients not responding to causative drug withdrawal. The treatment with TPE or RTX should be considered in immune-mediated DITMA, as described in association with quinine, gemcitabine, cisplatin and clopidogrel (Jacob et al., 2012; Padmanabhan et al., 2019). Both TPE and RTX lead to removal of drug dependent-antibodies, the first one via mechanical removal of autoantibodies (Nguyen et al., 2019; Jung et al., 2022) and the ladder permitting apoptosis of autoantibody producing cells as in other forms of TTP (Au et al., 2007; Jestin et al., 2018; Jung et al., 2022). TPE has been applied to immune -mediated DITMA based on the former positive experience with TTP, although there is no data on TPE efficacy or TPE superiority compared with other techniques (such as plasma infusion). Retrospective and prospective studies on large DITMA cohorts who underwent RTX therapy are missing. In relation to TPE, to the best of our knowledge, only retrospective cohort studies are reported, mainly for gemcitabine, clopidogrel and quinine. In gemcitabine-related TMA, patients treated with TPE showed a remission rate and prognosis similar to patients treated with only drug suspension, although in presence of more severe kidney and hematological manifestations at baseline (Gore et al., 2009; Daviet et al., 2019). In clopidogrel-associated TMA, similar remission rates are reported in both TPE and non TPE-treated groups (Bennett et al., 2007; Jacob et al., 2012). In quinine-associated TMA, comparative data on different remission rates between TPE and non TPE-treated groups are not reported because all recruited patients were treated with plasma-exchange (Page et al., 2017). Retrospective studies are also reported for TPE in ticlopidine-associated TMA (Bennett et al., 2007; Jacob et al., 2012). Although there are numerous concerns on the association between ticlopidine and TMA (due to temporal criteria), hematological and renal remission rates in TPE-treated patients are significantly better compared to non TPE-treated patients (Jacob et al., 2012; Bennett et al., 2013), supporting the possible involvement of drug-induced antibodies (with reduced anti-ADAMTS13 activity) in determining the ticlopidine-induced clinical scenario similar to TTP.

The utility of TPE in other etiologies of DITMA is debatable. TPE would be expected to be effective in immune-mediated DITMA, whereas it would not be with those medications in which direct endothelial injury occurs (Winters, 2017). However, drugs described to have direct endothelial toxicity induce release of large amounts of von Willebrand factor multimers and concomitant activation of the coagulation cascade. The theoretical concept of TPE in DITMA combines two main aspects: removal of harmful circulating molecules released during the endothelial damage (such as vWF multimers, proinflammatory cytokines and permeability factors) that directly contribute to the underlying etiopathogenetic mechanism, and replacement of physiologic plasma proteins (e.g., activated protein C, antithrombin, tissue factor pathway inhibitor, vWF cleaving proteases, VEGF, complement regulator factors) to ultimately restore hemostasis. In patients with DITMA, in association with TPE, corticosteroids (intravenous and/or per os) are used, but data on treatment efficacy are missing. Considering the underlying etiopathogenetic mechanism, their use could find application in immune-mediated DITMA.

Based on data currently available in the literature and considering the limited evidence, we suggest that the use of TPE and RTX in DITMA must be weighed and limited to a specific subset of patients with suspect of immune-mediate DITMA, in particular in the case of a possible association with ticlopidine, and in non-responsive to drug withdrawal (Izzedine and Perazella, 2015; Schwartz et al., 2016; Brocklebank et al., 2018; Padmanabhan et al., 2019). Moreover, according to consensus opinion (Palma et al., 2021), regardless of whether a toxic or immune-mediated form, a trial of TPE is recommended in all patients with a suspected diagnosis of severe DITMA, in addition to suspected-causative drug suspension. During TPE, platelet count, haemolysis lab tests and renal function should be carefully monitored. In summary, other studies are needed to confirm a potential role of TFE and RTX in DITMA.

Given the role played by complement system activation, anti-complement drugs have been proposed as a therapeutic option in DITMA. Eculizumab is a monoclonal drug that inhibits C5-factor, thus not permitting the formation of the Complement Membrane Attack Complex, avoiding, the facto, cell lysis and death.

Eculizumab is approved for the treatment of Paroxysmal Nocturnal Hemoglobinuria and aHUS, although it is widely adopted for patients with severe secondary TMAs (Caravaca-Fontan and Praga, 2019), even if its efficacy has not been demonstrated in validated clinical trials. The overactivation or deregulation of complement system is not proven in DITMA patients and, consequently, indications for eculizumab treatment are limited. Some Authors (Cavero et al., 2017; Caravaca-Fontan and Praga, 2019) suggest to use the anti-complement therapy only in case of lack of improvement of hematological parameters and/or renal function recovery after causative drug discontinuation. This approach lacks supporting evidence, which is currently based on case reports and case series. Other Authors (Duineveld and Wetzels, 2019)do not suggest the use of eculizumab in secondary DITMA cases, relying on data indicating that renal outcome was not significantly altered by the use of the complement inhibitor. Several case report and little case series suggested the effectiveness of C5-inhibitor in DITMA in terms of normalization of hematological parameters and, in some cases, partial or complete recovery of kidney function (Allinovi et al., 2017; Cavero et al., 2017; Caravaca-Fontan and Praga, 2019; Grall et al., 2021). In particular, eculizumab proved its efficacy in three out of 9 (33.3%) patients who resolved the IFN-related TMA without further complications, and in four out of six patients who needed dialysis treatment at DITMA onset and discontinued dialysis after a mean of 2 months of eculizumab treatment (Allinovi et al., 2021). Moreover (Grall et al., 2021), reported a significantly better kidney response (83% vs. 64% of complete/partial recovery) and kidney outcome (eGFR 45 vs. 33 ml/min/1.73) in 13 patients with gemcitabine-induced TMA treated with anti-complement therapy. However, conflicting results were reported by previous articles gemcitabine-induced TMA treated with C5-inhibitor (Al Ustwani et al., 2014; Daviet et al., 2019). Data on large DITMA cohorts and on other causative DITMA drugs are not available.

As we reported in the previous paragraph, decreased C3 fraction in blood and/or deposition of C3 fraction in kidney biopsy may be considered as “red flags” of an overactivation of complement system, on which the choice to start a treatment with a complement inhibitor should be focused.

In this review, considering that DITMAs rarely conceal primitive TMAs (2/122, 1.6%), we support the flow-chart of TMA treatment previously proposed (Brocklebank et al., 2018; Caravaca-Fontan and Praga, 2019) in which the removal of the trigger should be pursued in suspected secondary TMAs (i.e. withdrawal of the offending drug in DITMA). The use of eculizumab in DITMA is suggested in severe TMA cases, and when the drug withdrawal and 3–5 TPE sessions do not associate with an improvement in renal function and hematological parameters. In these cases, complement-mediated TMA may need to be considered as an alternative diagnosis. In fact, it is difficult to discriminate the etiology (primary vs. secondary) in the acute phase of TMA due to different reasons:

- Primary TMA is suspected based on the absence of “secondary causative factors” such as drugs, cancer, infection or pregnancy. However, primary TMA diagnosis is excluded only by negative genetic analysis, and “secondary causative factors” can frequently trigger primary TMA cases. Moreover, timing for the genetic analysis of complement factors generally takes 1–6 months, an aspect that does not fit with the acuteness and severity of TMA;

In conclusion, the efficacy of eculizumab treatment in DITMA needs to be clarified with further prospective studies, with homogeneous timing of eculizumab infusion and selection of patients with similar clinical and prognostic features.